Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!news.mira.net.au!inquo!news.uoregon.edu!cyclic.gsl.net!news.gsl.net!newsfeed.nacamar.de!news.nacamar.de!uunet!in3.uu.net!157.175.111.1!iphase.com!usenet From: Dave Littell <dlittell@iphase.com> Newsgroups: comp.unix.bsd.freebsd.misc Subject: The end of 2.1? Date: Tue, 18 Mar 1997 16:32:45 -0600 Organization: Interphase Corporation Lines: 104 Message-ID: <332F180D.16E3@iphase.com> NNTP-Posting-Host: @sw22-fddi.iphase.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; I; SunOS 5.5.1 sun4c) Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:37278 First off, I'd like to express my deep appreciation to the Core Team as well as all of the supporters of FreeBSD. You've all done (and are still doing) a tremendous job and performed quite a miracle in making PC hardware seem like a real computer. Having said that, we now turn the coin over... While I understand that most of the development effort is now directed towards 2.2 and 3.0 and that maintenance activities are in general no fun at all, there are a number of what I consider very serious problems remaining in 2.1. I committed to 2.1 primarily because of the billing "rock-stable" and so on. While 2.1 has been infinitely better than any Microshit product ever could be, I've run into a number of problems that could be highly visible (and very disconcerting) to my management and potential customers. Following is a general list of things that are/have been worrisome in my 2.1.5/6/7-based development efforts: - There's some kind of problem between 2.1.6 and the new Solaris 2.5.1 DNS. After our Lab guys switched over to the new DNS, NFS mounts hang and never complete. Switching back to the previous DNS allows mounts to complete successfully. Other machines using the new DNS can mount with no problems. - The latest and greatest NFS security problem: will this in itself spur another 2.1 release? (This is sort of a rhetorical question, keep reading.) - The previously discussed keyboard lockup problem manifests itself more often than I'd really like (which is never). - The various ongoing battles with PPP - I'd love a 2.1 release with all the fixes. - DHCP: I need something solid in 2.1. - Gazing over the list of critical problems doesn't exactly instill a deep sense of peace. Jeez, a filesystem problem over 2 years old (kern/216)? - Endless install problems with the ports collection. Install attempts either do nothing (gnat) or fail to compile due to some ridiculous __FreeBSD_version "#if" (too many to list). Who knows what else lurks outside of my normal usage path? I've been very pleased with the operational quality the ports (once I finally get them installed). That is, with the ones I haven't completely abandoned. The point of all this is there needs to be a LOT more quality control on the ports collection. If I'm to continue to try to use, develop with, and deploy 2.1 (or indeed, any release) I simply can't have filesystem corruptions, keyboard lockups, endless PPP patches, and other red-flag problems. I'd like to think that anyone developing commercial products would agree. While I am acutely/painfully aware of the "need" for aggressive marketing, I am concerned by the claims that the ports collection contains applications that are "ready-to-run". It's a battle just to get some of them to compile. I believe a full install of each of the ports should be attempted for every release of the OS. If they don't build and install correctly, they shouldn't be included in the ports collection for that release. Very simple, very straightforward, and at the very least avoids the more obvious silliness (like the __FreeBSD_version bullshit I've run into time and again in attempting to install ports under 2.1.6). What I'd like to grovel, plead, and beg for is an organized, focused, and serious effort towards fixing the remaining known problems in 2.1 and issuing enough maintenance CD releases to ultimately get a 2.1.x that is truly "rock-stable". Yeah, I can continue to hack and patch and cobble together ad infinitum, but being forced to do so doesn't give me the credibility I'd like (and need) when I go to my management and try to sell FreeBSD as the OS of choice for my projects. I also think it would be helpful if the Release Notes indicated fixes/changes in more specific terms than what's currently provided. Now that I've stomped on nearly everyone's toes and pissed in all the sacred sandboxes, I'm sure I've made quite a few new enemies here. As you fuel your flamethrowers, just bear in mind I'm just pointing out what (in my opinion) are issues that must be effectively addressed as FreeBSD struggles to move beyond the toy/project domain and into the real world. Whaddya think? Dave Disclaimer: Everything expressed here is strictly my personal opinion. Deal with it.